Polydnavirus Integration in Lepidopteran Host Cells in Vitro D.E

Journal of Insect Physiology 49 (2003) 453–462 www.elsevier.com/locate/jinsphys

Polydnavirus integration in lepidopteran host cells in vitro D.E. Gundersen-Rindal ∗, D.E. Lynn US Department of Agriculture, Agricultural Research Service, Insect Biocontrol Laboratory, Beltsville, MD 20705, USA

Received 14 July 2002; received in revised form 11 October 2002; accepted 14 October 2002

Abstract

The long-term persistence of polydnavirus (PDV) DNA in infected lepidopteran cell cultures has suggested that at least some of the virus sequences become integrated permanently into the cell genome. In the current study, we provide supportive evidence of this event. Cloned libraries were prepared from two different Lymantria dispar (gypsy moth) cell lines that had been maintained in continuous culture for more than ﬁve years after infection with Glyptapanteles indiensis PDV (GiPDV). Junction clones containing both insect chromosomal and polydnaviral sequences were isolated. Precise integration junction sites were identiﬁed by sequence comparison of linear (integrated) and circular forms of the GiPDV genome segment F, from which viral sequences originated. Host chromosomal sequences at the site of integration varied between the two L. dispar cell lines but virus sequence junctions were identical and contained a 4-base pair CATG palindromic repeat. The GiPDV segment F does not encode any self-replication or self-insertion proteins, suggesting a host-derived mechanism is responsible for its in vitro integration. The chromosomal site of one junction clone contained sequences indicative of a new L. dispar retrotransposon, including a putative reverse transcriptase and integrase located upstream of the site of viral integration. A potential mechanism is proposed for the integration of PDV DNA in vitro. It remains to be seen if integration of the virus also occurs in the lepidopteran host in vivo. Published by Elsevier Science Ltd.

Keywords: Integration; Transformation; Retrotransposon; Cell culture; Polydnavirus; Glyptapanteles indiensis; Lymantria dispar

1. Introduction 1986; Lavine and Beckage, 1995; Lawrence and Lanzr- ein, 1993; Stoltz, 1993; Summers and Dibb-Hajj, 1995; Polydnaviruses (PDVs), unique mutualistic insect Strand and Pech, 1995), presumably under the control viruses found in some parasitic Hymenoptera, are of larval host factors that are not well understood. No characterized by their unusual multi-segmented closed response is mounted to eliminate the parasitoid eggs and circular dsDNA genomes. PDVs are considered as non- they are allowed to develop, hence PDVs play essential traditional viruses because their replication has been roles in the parasitoid life cycle (Stoltz, 1993). detected only in oviduct calyx cells within the female The two recognized PDV genera, ichnoviruses (IVs) parasitoid reproductive tract. PDVs are injected along and bracoviruses (BVs) associated with parasitoid wasps with eggs by the parasitic wasps into their lepidopteran of Ichneumonidae and Braconidae, respectively, differ in host hemocoel during oviposition. Inside the larval host, their particle morphology, physiology, molecular charac- PDVs infect cells but do not replicate. Their genomes are teristics, and host range (Stoltz and Whitﬁeld, 1992; evidently transcriptionally active (Blissard et al., 1986; Stoltz et al., 1995; reviewed in Webb, 1998). Genetic Strand et al., 1992) within the host, and their transiently studies have indicated that both IVs and BVs share the expressed gene products function to suppress the normal property of vertical transmission through the germ line immune activity and regulate development and physi- (Stoltz, 1990; Stoltz et al., 1986), through integration ology of the host (Asgari et al., 1997; Blissard et al., into their respective wasp host genomes as proviruses (Fleming and Summers, 1990; Fleming and Summers, 1991; Fleming, 1991; Gruber et al., 1996; Savary et al., ∗ Corresponding author: Tel.: +1-301-504-6692; fax: +1-301-504- 5104. 1997; Stoltz, 1990; Stoltz, 1993; Stoltz et al., 1986; E-mail address: [email protected] (D.E. Gundersen- Wyder et al., 2002). Evidence for transmission through Rindal). proviral genome integration has been shown for IVs

0022-1910/03/$ - see front matter Published by Elsevier Science Ltd. doi:10.1016/S0022-1910(03)00062-3 454 D.E. Gundersen-Rindal, D.E. Lynn / Journal of Insect Physiology 49 (2003) 453–462 associated with Campoletis sonorensis (the IV type several GiPDV-encoded genes in long-term transformed species) (Fleming and Summers, 1986; Fleming and LdEp cells appear to be transcribed (Gundersen-Rindal, Summers, 1990) and Hyposoter fugitivus (Xu and Stoltz, unpublished results). 1991) and for BVs associated with Chelonus inanitus With the exception of one site from GiPDV/LdEp (Gruber et al., 1996; Wyder et al., 2002) and Cotesia (Gundersen-Rindal and Dougherty, 2000), neither PDV congregata (Savary et al., 1997). The viral sites of pro- in vitro integration border sites containing virus and viral integration are marked by terminal repeats, which chromosomal sequence have been isolated nor potential vary in length and homology with each PDV studied, means or mechanisms for integration examined, for any and markedly between IVs and BVs (Fleming and Sum- PDV systems. Because our laboratory is interested in mers, 1991; Cui and Webb, 1997; Gruber et al., 1996; virus-mediated in vitro cell transformation, chromo- Wyder et al., 2002). PDV circular genome segments somal and viral junction sequences of a GiPDV circular appear to be produced by excision at these repeats from genome segment integrated in vitro in two different L. the parasitoid genome, by a mechanism similar to those dispar cell lines were isolated and analyzed to investi- used by transposable elements, retroviruses, and retro- gate the nature of PDV in vitro integration. transposons (reviewed in Webb, 1998). Studies concerning the fate of PDVs within their natu- ral lepidopteran hosts have primarily focused on virus- 2. Materials and methods specific gene expression and the physiological changes associated with parasitism (Stoltz, 1993) and have been 2.1. GiPDV-transformed L. dispar cell lines conducted in vivo. A few studies have examined the fate of PDVs and the effects of PDV infection in vitro. To L. dispar cell lines originally derived from embryonic date, three PDVs have been described and shown to per- tissue (IPLB-LdEp) and pupal ovaries (IPLB-Ld652Y) sist over time, maintained in vitro in infected cell popu- were infected with GiPDV in vitro as previously lations derived from lepidopterans. Two are the IVs described (McKelvey et al., 1996). Total genomic associated with Hyposoter fugitivus (HfPV) and Hypo- nucleic acids were extracted from non-infected and soter didymator (HdIV). HfPV was used to infect a cell GiPDV-infected IPLB-LdEp (IPLB-LdEp/Gi) and IPLB- line derived from Lymantria dispar (Ld652Y), a non- Ld652Y (IPLB-Ld652Y/Gi) cells, after greater than 250 permissive host of the parasitoid (Kim et al., 1996). passages, by standard techniques (Ausubel et al., 1994). HdIV was used to infect several cell lines derived from host Spodoptera littoralis haemocytes, as well as non- 2.2. Isolation of GiPDV DNA and GiPDV genomic preferred hosts of the parasitoid Spodoptera frugiperda segment clones (SF9) and Trichoplusia ni (High Five) (Volkoff et al., 1999, 2001). The third is the BV associated with Glypta- Female braconid G. indiensis parasitic wasps were panteles indiensis (GiPDV) in several cell lines derived dissected and calyx fluid containing PDVs extracted for from host L. dispar (McKelvey et al., 1996; Gundersen- nucleic acid as previously described (Beckage et al., Rindal and Dougherty, 2000) and non-preferred hosts of 1994). A GiPDV plasmid clone library was generated the parasitoid, including several lepidopteran (T. ni, S. using the method described by Albrecht et al. (1994) to frugiperda, Plodia interpunctella, and Heliothis isolate complete GiPDV circular genomic segments. The virescens) and a coleopteran (Diabrotica library was screened to identify GiPDV circular seg- undecimpunctata) cell lines (Gundersen-Rindal et al., ments, including p157 analyzed in this study, capable of 1999). In the examples of persistent HfPV and GiPDV persisting in infected L. dispar cell in vitro as previously infection, cytopathic effects were observed in cells described (Gundersen-Rindal and Dougherty, 2000). The initially upon infection and for several weeks, followed full ds circular segment represented by clone p157 was by a period of cellular recovery in which the cell popu- recently designated as segment F of the GiPDV genome lation morphology and growth patterns returned essen- (Chen and Gundersen-Rindal, in press) and is referred tially to normal. In all three examples, viral sequences to as segment F/p157 throughout. were maintained in the recovered and apparently transformed cell lines in association with high molecular 2.3. Restriction analysis of GiPDV-transformed L. weight cellular DNA of presumed chromosomal origin. dispar cell DNAs for GiPDV sequences persisting in Furthermore, the viral sequences were readily detectable vitro after 18 up to 36 months, even after 60 months for HdIV/SF9 and GiPDV/LdEp, of routine cell subculture. Digoxigenin-labeled plasmid DNA from GiPDV seg- For HdIV, an uncharacterized K-gene was shown to be ment F/p157 was used as a hybridization probe to ana- transcribed and expressed in long-term transformed Spo- lyze transformed IPLB-LdEp/Gi cell DNA for evidence doptera and Trichoplusia cells, while other HdIV genes of integrated GiPDV viral sequences. For Southern blot, were not transcribed (Volkoff et al., 2001). Likewise, cloned segment F/p157 (1ug) and IPLB-LdEp/Gi (5 ug D.E. Gundersen-Rindal, D.E. Lynn / Journal of Insect Physiology 49 (2003) 453–462 455 each) DNAs were digested separately with restriction (integrated) of IPLB-LdEp/Gi clone p7-54 and IPLB- endonucleases BamHI, KpnI, PstI, SstI, and XbaI Ld652Y/Gi clone L16XE11 were compared with GiPDV (Invitrogen, Carlsbad, CA), electrophoresed through a sequences (circular) obtained from GiPDV circular DNA 0.8% agarose gel, transferred to nylon membrane segment F/p157 and integration junctions identified. All (Southern, 1975), and probed. Uncut GiPDV (0.5 ug) sequences were analyzed for homologies to known and IPLB-LdEp (5 ug) genomic DNAs were included as sequences in GenBank using BLAST (Altschul et al., positive and negative controls, respectively. The tem- 1990, 1997). Sequences for junction clones p7-54 perature for hybridization was 48 °C. Annealed DNAs (updated sequence) and L16XE11 have been deposited were visualized using the chemiluminescent detection in GenBank under accession numbers AF198385 and reagent disodium 2-chloro-5-(4 methoxyspiro [1,2-diox- AY162267, respectively. etane-3,2Ј-(5Ј-chloro)-tricyclo [3,3,1,13,7]decan]- 4yl)phenyl phosphate)(CSPD) (Roche Biochemicals, Indianapolis, IN). GiPDV segment F/p157 was 3. Results sequenced by a combination of primer walking and using the GPS-1 genome priming system (New England Biol- 3.1. GiPDV-transformed L. dispar cell lines abs, Beverly, MA) on an ABI310 automatic sequencer (Applied Biosystems, Foster City, CA). A restriction Infection of two L. dispar cell lines derived from map based on restriction endonuclease recognition sites embryonic tissue with GiPDV in vitro was previously was generated from the sequence using the MapDraw described (McKelvey et al., 1996). The GiPDV-infected component of DNASTAR software package (Madison, IPLB-LdEp (IPLB-LdEP/Gi) and Ld652Y (Ld652Y/Gi) WI). The putative site of viral integration for this seg- cells utilized in these experiments had been routinely ment was located on the map. passed in the laboratory at least 250 times when genomic DNA was extracted for libraries, indicating that the 2.4. Isolation and analysis of chromosomal:viral GiPDV viral sequences were stably maintained as a per- integration junctions manent part of the replicating cells and the cells were, in fact, transformed. Two genomic libraries were derived from GiPDV- infected Ld cells. The IPLB-LdEp/Gi clone p7-54, con- 3.2. GiPDV DNA segment maintained in infected taining an integration junction, was derived by subclon- transformed Ld cells in vitro ing from an IPLB-LdEp/Gi lambda phage library as described previously (Gundersen-Rindal and Dougherty, Clone p157 was 18.6 kb in size as determined by 2000). The IPLB-Ld652Y/Gi library was derived by restriction and full sequence analysis and represented a ligating BamHI partially-digested DNA, filled in using complete dsDNA circular segment of the GiPDV gen- the Klenow fragment of DNA polymerase I, into the ome as previously described. In characterization of the XhoI site of vector Lambda Fix II (Stratagene, Cedar GiPDV genome, the viral segment represented by clone Creek, TX). Clones hybridizing to digoxigenin-labeled p157 was designated as GiPDV genome segment F GiPDV were isolated by plaque lift. Clones were ana- (Chen and Gundersen-Rindal, submitted). Segment F is lyzed by restriction digestion followed by Southern presented graphically (Fig. 1A) in the orientation in hybridization using digoxigenin-labeled total GiPDV which it was originally cloned and sequenced. The same and IPLB-Ld652Y DNAs separately as probes. Lambda restriction profile is generated by analysis of the reverse clones containing fragments that hybridized with both complement or opposite orientation of the circle, and the viral and cellular probes were sub-cloned by inserting full circular segment F has been cloned in both orien- either HindIII or XbaI fragments into the pGEM3zf(+) tations (data not shown). plasmid vector (Promega Corp., Madison, WI). Sub- Each PDV-infected insect cell line examined to date clones were screened by reciprocal hybridization using for in vitro persistence of polydnaviral sequences has labeled GiPDV and IPLB-Ld652Y DNAs as probes. indicated that part, but not all, of the viral genome is Several sub-clones containing both viral and chromo- maintained over time. However, no studies have exam- somal sequences were isolated and analyzed. Clone ined in detail which viral genome segments or which Ld16XE11 from this library, as well as p7-54 from the regions of segments are maintained. GiPDV-infected IPLB-LdEp/Gi library, were sequenced after using the transformed cell line IPLB-LdEp/Gi was examined for GPS-1 genome priming system (New England Biolabs) the persistence of segment F/p157 sequences in order to on an ABI310 automatic sequencer (Applied analyze which sequences were maintained over time in Biosystems) and determined to originate from the same vitro. Restriction endonuclease recognition sites for PDV genome segment. Sequences were assembled using BamHI, XbaI, SstI, PstI, and KpnI were mapped on the the SeqManII component of the DNASTAR software segment F/p157 circle shown in Fig. 1A. Corresponding package (Madison, WI). The GiPDV sequences digests of the cloned segment F/p157 circle were shown 456 D.E. Gundersen-Rindal, D.E. Lynn / Journal of Insect Physiology 49 (2003) 453–462

Fig. 1. In vitro maintenance of GiPDV circular ds DNA genome segment F sequences, cloned in its entirety as p157, within infected IPLB- LdEp/Gi cells. (A) Restriction map of GiPDV segment F/p157, generated from full nucleotide sequence of clone p157 (including the plasmid vector sequence) using the MapDraw component of DNASTAR software package. Fragments that were identified within in vitro transformed IPLB- LdEp/Gi chromosomal DNA with all or most restriction endonucleases are shaded black and gray, respectively. The white shaded regions represent KpnI and other restriction profiles indicating the presence of the entire segment F/p157 circle within IPLB-LdEp/Gi chromosomal DNA. Segment F/p157 fragment sizes (in kilobase) predicted are: BamHI, 11.2,7.1, & 3.2; XbaI, 13.0 & 8.6, SstI, 10.9, 5.8, 3.3, &1.5; PstI, 8.8, 6.0, 3.6 & 3.2; KpnI, 21.6. (B) Southern blot analysis of transformed GiPDV-infected insect cell line IPLB-LdEp/Gi chromosomal DNA for persisting viral sequences. Uncut genomic DNA from GiPDV and non-infected cells, and DNAs digested separately with restriction endonucleases BamHI, XbaI, SstI, PstI, and KpnI from clone p157 (GiPDV segment F) (a lanes) and IPLB-LdEp/Gi cell line (b lanes) were electrophoresed through a 0.8% agarose gel and transferred to nylon membrane. Digoxigenin-labeled clone p157 (GiPDV segment F) DNA was used as a hybridization probe. Lane S contains molecular size standard fragments of kilobase sizes as labeled. in lane a for each restriction endonuclease. Digests of (Gundersen-Rindal and Dougherty, 2000) and was IPLB-LdEp/Gi chromosomal DNAs were shown in lane refined in this study (see Fig. 1B). Most IPLB-LdEp/Gi b, after hybridization with the labeled cloned segment restriction profiles showed the absence of the segment F/p157 probe. Comparison of IPLB-LdEp/Gi restriction F/p157 restriction fragment containing the putative site profiles with those of the mapped segment showed that of integration (Fig. 1B). For example, the 11.2 kb BamHI most to all of segment F/p157 was maintained in fragment containing the putative site of integration is infected cells. Several IPLB-LdEp/Gi restriction profiles absent, and its viral sequences presumably may contrib- (Fig. 1B, BamHI, SstI, and PstI) indicated the presence ute to the visible undefined higher molecular weight of higher molecular weight bands that were not well fragments. defined. These were suggestive of fragments that might have been generated from endonuclease digestion at one 3.3. Sequence analysis of the junction sites of viral and one chromosomal recognition site, often integrated GiPDV DNA referred to as off-size fragments. These signs of viral integration were potentially generated at more than one From a previously isolated and sequenced border chromosomal locus, but were not seen in association clone from transformed IPLB-LdEp/Gi cells, the with every restriction endonuclease profile. Restriction sequence of a single putative site of in vitro integration fragments present in IPLB-LdEp/Gi were mapped to the of segment F/p157 was identified. To analyze events sur- circular segment. The regions of the segment F/p157 cir- rounding in vitro integration of GiPDV in Ld cells, cle highlighted in black or gray on Fig. 1A identify frag- additional junction sites were needed. The clone p7-54, ments that were identified within in vitro transformed which originated from a lambda phage clone containing IPLB-LdEp/Gi chromosomal DNA with the majority of both viral and host chromosomal sequences and con- restriction endonucleases. Other restriction profiles indi- tained the LdEp/Gi putative integration junction cated the presence of the entire segment F/p157 circle described previously, was sequenced to examine the (white shaded on Fig. 1A) within IPLB-LdEp/Gi chromosomal sequences in more detail. A second library chromosomal DNA, for example KpnI(Fig. 1B) and cloned from transformed IPLB-Ld652Y/Gi cells and EcoRI (data not shown). These restriction sites occurred screened to isolate junction clones yielded several, in close proximity to the site of viral integration. among them clone L16XE11. Both L. dispar (IPLB- The putative site of segment F/p157 integration in LdEp and IPLB-Ld652Y) junction clones contained infected IPLB-LdEp cells had been identified previously GiPDV sequence originating from the GiPDV segment D.E. Gundersen-Rindal, D.E. Lynn / Journal of Insect Physiology 49 (2003) 453–462 457

F/p157. We identified the precise integration junction estingly, the identified viral sites of integration were between GiPDV and its flanking chromosomal identical in both L. dispar cell lines. GiPDV sequences sequences by comparing sequence of the circular and immediately at the integration junction were “CATG” in integrated (p7-54 and L16XE11) forms (Fig. 2A). Inter- both. Surprisingly, the L. dispar chromosomal borders

Fig. 2. Sequence analysis of integration junctions obtained from two GiPDV-infected Lymantria dispar cell lines. (A) Chromosomal:viral junction sites determined by comparing circular and integrated forms of GiPDV segment F/p157 from GiPDV-infected L. dispar cell lines IPLB-LdEp/Gi (clone p7-54) and IPLB-Ld652Y/Gi (clone L16XE11). Both strands of GiPDV segment F DNA sequences at integration junction site are shown. GiPDV viral sequences are italicized. Both clone sequences are given in the reverse complement orientation from which they were cloned. (B) Chromosomal and viral sequences of IPLB-LdEp/Gi junction clone p7-54. Integration site is indicated by an arrow. GiPDV viral sequence are italicized. (C) Chromosomal and viral sequences of IPLB-Ld652Y/Gi junction clone L16XE11. Integration site is indicated by an arrow. GiPDV viral sequence are italicized. Analysis using BLAST of L. dispar chromosomal sequences indicated a small region with 3Ј to 5Ј nucleotide homology to sequences of the Bombyx mandarina Pao-like retrotransposon Yamato, underlined, and an ORF with amino acid homology to the pol protein of Bombyx mori, encoding reverse transcriptase and integrase, upstream of the site of integration (amino acid translation is shown). The integrase conserved cysteine-histidine motif is boxed. 458 D.E. Gundersen-Rindal, D.E. Lynn / Journal of Insect Physiology 49 (2003) 453–462 did not have sequence identity or any identifiable com- moth, L. dispar, to date, though they are known to exist mon motif. Studies of PDV integration in the respective in nearly every insect and eukaryotic system. wasp host chromosome as a provirus shows that inverted GiPDV segment F/p157, other GiPDV genome seg- repeated sequences (sometimes lengthy) are present at ments, or other polydnavirus genome segments integration junction sites (Savary et al., 1997). In this sequenced to date do not encode integrase, transposase, example of GiPDV chromosomal integration, palin- polymerase, or other protein(s) that could facilitate self dromic repeated sequences “CATG,GTAC” were ident- chromosomal insertion in vitro. Furthermore, excision ified directly at the site of in vitro integration. Viral enzymes are unknown for polydnaviruses. The pol gene sequences of segment F/p157 were characterized by of retroviruses and retrotransposons encodes a reverse numerous A and T or AT runs. transcriptase followed by an integrase (IN protein) at the L. dispar chromosomal sequences in proximity to the 3Ј end. The pol gene of retroviruses and retrotransposons site of viral integration in LdEp/Gi clone p7-54 (Fig. 2B) usually contains a conserved cysteine-histidine motif. In were comparatively unremarkable. Sequence analysis retroviruses, the integrase domain following the reverse using BLAST (Altschul et al., 1990, 1997) revealed transcriptase domain usually contains a HX3HX22– incomplete or limited nucleotide homologies to known 32CX2C motif (Johnson et al., 1986) while a CXCX8– genes, for example a 61-nucleotide homology to the vit- 9HX4C integrase motif is generally conserved among ellogenin gene of L. dispar. While clearly chromosomal divergent species of non-LTR retrotransposons in origin, this locus was essentially non-coding, parti- (Jakubczak et al., 1990). In our example of PDV inte- cularly in the region upstream of viral integration. gration in vitro, junction clone L16XE11, L. dispar retrotransposon-like chromosomal sequences encoded an ORF for a pol-like protein in which the CXCX HX C 3.4. Structural similarity of Ld652Y viral integration 8–9 4 motif characteristic of non-LTR retrotransposons was site to retrotransposon present but the spacing between the first C and second C (CXC) was changed to an interval of two bases

A longer stretch of host chromosomal nucleotide (CX2C). This same interval change has been reported sequence was obtained within IPLB-Ld652Y/Gi junction for SART1 non-LTR retrotransposons of Bombyx mori clone L16XE11 (Fig. 2C) and L. dispar sequences in (Takahashi et al., 1997). proximity to the site of viral integration in Ld652Y were Based on our analysis of GiPDV segment F/p157 inte- more informative. Analysis using BLAST of these L. gration into two different L. dispar cell lines in vitro, a dispar chromosomal sequences indicated a small region putative integration model is proposed (Fig. 3) in which (underlined) with 3Ј–5Ј nucleotide homology to the palindromic CATG on segment F is recognized and sequences of the Bombyx mandarina Pao-like retrotran- the circular viral segment is integrated via host-encoded sposon Yamato (GenBank accession AB055223) (Abe reverse transcriptase/integrase. et al., 2001), and an open reading frame (ORF) with amino acid homology to the pol protein of Bombyx mori (GenBank accession T18196) (Takahashi et al., 1997), 4. Discussion encoding a reverse transcriptase and integrase, located just upstream of the chromosomal site of integration Over time, parasitoid wasps have evolved mutualistic (Fig. 2C). These homologies pointed to a potential L. relationships with polydnaviruses. The entire virus gen- dispar retrovirus or retrotransposon. Retroviruses are ome is integrated within the wasp genome as provirus RNA viruses that are converted to a dsDNA form by and transmitted vertically, as part of the wasp’s genetic reverse transcriptase prior to integration at random sites material, to progeny. PDVs share some characteristics into the host cell chromosome via an integrase, the pro- with transposable elements in that PDV circular genomic tein responsible for integration of viral DNA into host segments are integrated within the parasitoid genome genome(s) (Grandgenett, 1986). Retrotransposons are and derived by excision from these sites at the onset mobile DNA elements within eukaryote genomes that of viral replication. Prior to excision, integrated proviral transpose via RNA intermediates using self-encoded segments are characteristically marked by terminal reverse transcriptase and integrase. Retrotransposons repeated sequences that vary considerably in length, have been classified into two groups according to struc- homology, and complexity (Fleming and Summers, ture, whether they contain long terminal repeats (LTRs) 1991; Gruber et al., 1996; Cui and Webb, 1997; at both ends (LTR type or retrovirus-like element) or not reviewed in Webb, 1998; Wyder et al., 2002). The pro- (non-LTR type element) (Eickbush, 1992; Gabriel and viral C. sonorensis ichnovirus (CsPDV) segments appar- Boeke, 1993; Hutchison et al., 1989). Retrotransposons ently are integrated at multiple loci, flanked on both sides may be integrated at random or at specific sites on the by parasitoid DNA (Fleming and Summers, 1991; Cui chromosome. No retroviruses or retrotransposons have and Webb, 1997). In contrast, Chelonus inanitus braco- been identified specifically in association with the gypsy virus (CiV) segments appear clustered and present in D.E. Gundersen-Rindal, D.E. Lynn / Journal of Insect Physiology 49 (2003) 453–462 459

Fig. 3. Putative model for the integration of polydnavirus DNA into insect cell chromosomal DNA as described in the text, suggested by examination of chromosomal:viral junction sites in Ld cells. The sequences “CATG” and its palindrome “GTAC” are represented by heavy black lines, with the joined CATGGTAC double length. tandem array, flanked on both sides by virus DNA dence shows that DNA from at least three different cir- (Wyder et al., 2002), as does the EP1 segment of braco- cular genome segments is maintained in vitro. In this virus Cotesia congregata (CcPDV), flanked on one side study, chromosomal maintenance of a single GiPDV cir- by parasitoid and the other side by viral segment DNA cular segment, segment F, cloned in its entirety as p157 (Savary et al., 1997). Unlike transposable elements, PDV as described previously, was examined in detail. This segment excision occurs only in certain tissues at spe- segment is the best characterized of the GiPDV genome, cific later stages of development, and evidence from at and was previously shown by this laboratory to be main- least two PDV systems suggests that segments do not tained in vitro in transformed IPBL-LdEp cells with evi- re-integrate into chromosomal DNA (Theilmann and dence strongly supporting chromosomal integration. Summers, 1986, 1987; Strand et al., 1992). Excised PDV Here we have shown by mapping experiments that at segments are circularized to form the PDV genomic seg- least a large part of the segment F/p157 circle, in which ments, which, after release from ovarian calyx cells, are restriction fragments containing the site of viral inte- injected into parasitized larvae during oviposition to gration are “missing” and may contribute to off-size cause host immune-suppression and developmental fragments, is consistently maintained chromosomally. delays in a manner that favors parasitoid survival. The The evidence suggests that the entire GiPDV segment mutualism between virus and wasp is so complete and F/p157 is maintained in vitro, with restriction profile ancient that one could consider the provirus segments as varying by proximity of the restriction recognition legitimate parasitoid genetic material (Whitfield, 2002). sequence to the site of integration. Several laboratories have recently described infection Our examination of sequences at the junction site of of insect-derived cell lines with PDVs, both IVs and a the same GiPDV circular segment (F/p157) in two dif- BV (Kim et al., 1996; McKelvey et al., 1996; Volkoff ferent L. dispar-derived cell lines has provided the first et al., 1999), and the apparent ability of some PDV evidence for a mechanism of PDV integration in vitro. sequences to persist in these cells in vitro over time. First, F/p157does not encode integrase, transposase, or The phenomenon of PDV in vitro integration has been any other known protein(s) that can facilitate self inser- puzzling because it has been unclear biologically why tion (nor do any of the other GiPDV genome segments such an event might occur, and, furthermore, whether the sequenced to date, Gundersen-Rindal, unpublished) sug- in vitro PDV infections might mirror in vivo processes gesting that a host-derived mechanism is being occurring in PDV-infected larvae. There seem to be employed. Also, the insertion on the host genome was numerous potential explanations for PDV sequences different between the two Ld lines, but the virus integrated in vitro. Among these, PDVs may integrate sequence junctions were the same in each. This suggests parts of their genome in a manner that interrupts or the actual host insertion site may be random but that a affects expression of critical host gene(s). Alternatively, specific GiPDV sequence is being recognized. A clue host cells may actively mediate acquisition of PDV for the mechanism also comes from our examination of sequences with resulting integration into their junction clone L16XE11, in which Ld652Y chromo- chromosome(s). While the biological significance of somal sequences adjacent to the site of integration were PDV persistence in vitro has not been at all clear, the characteristic of a L. dispar-associated retrovirus or potential to use PDVs for controlled transformation of retrotransposon. These sequences encoded an ORF for a insect cells has been appreciated. pol-like protein in which an integrase domain, CXCX8– For the PDVs examined in vitro to date, the persist- 9HX4C, conserved among non-LTR retrotransposons, as ence of some, but not all, of the large PDV DNA genome well as some group II introns, was identified. The pol in vitro has been noted. In the GiPDV system, our evi- gene of retroviruses and retrotransposons encodes two 460 D.E. Gundersen-Rindal, D.E. Lynn / Journal of Insect Physiology 49 (2003) 453–462 domains: a reverse transcriptase followed by an inte- fashion as a provirus. Furthermore, PDVs in general self- grase. Reverse transcriptase is able to act on DNA tem- encode very few proteins related to their replication, but, plate and initiate a double strand break. The integrase rather, seem to rely on host machinery (Webb, 1998). protein uses linear DNA as its template to form a pre- Thus, such a mechanism where virus is integrated using integration complex and on a 5ЈCA(OH) 3Ј, on a phos- host-encoded gene products (presumably, in this system, phodiester bond of DNA target, initiates a strand transfer L. dispar retrotransposon pol gene-encoded reverse reaction. One can envision a mechanism (Fig. 3) where transcriptase/integrase acting “in trans”) would fit well specific target site sequences on the GiPDV circular seg- within the unusual life cycle evolved by polydnaviruses. ment F/p157, namely “CATG”, are recognized by In addition, such a mechanism offers important clues reverse transcriptase/integrase and a double strand break into how polydnaviruses might disassemble and propa- initiated. This would be followed by the L. dispar retro- gate themselves during the cycle of infection. transposon-like element integrase acting “in trans” to The ability of all PDVs as a general phenomenon, mediate ds DNA joining to result in the linearized and both IVs and BVs, to infect and essentially permanently integrated GiPDV segment, with the 8-base pair transform insect cell populations in vitro has yet to be “CATGGATC” site duplicated on each end. investigated. If the phenomenon of integration of PDV Furthermore, the evidence supports the integration of sequences in vitro is related evolutionarily to the nature the full segment F/p157 in vitro in both Ld cell lines. of the virus it seems logical that some or all PDVs would We suggest the “CATGGTAC” immediately at the site exhibit a similar ability to persist in vitro. This remains of integration may represent a 4-base pair palindromic to be seen, as not all cell lines infected with PDVs in direct repeat sequence for this segment and a recognition vitro and examined for evidence of viral persistence have site for both recombination and in vitro integration. This shown evidence of chromosomal maintenance. GiPDV sequence could potentially represent the circularization remains the first and, to date, only BV shown capable point for the segment after its excision from the parasit- of in vitro integration. Since IVs and BVs do not share oid genome. Palindromic termini to LTR sequences have a common evolutionary origin (Whitfield, 1997, 2002), been identified in other PDV systems, in most detail by suggesting they were derived independently from differ- Cui and Webb (1997) in which the segment W of ent ancestral viruses, in vitro viral maintenance and other CsPDV identical 1.2 kb LTRs present at the site of pro- features sometimes exhibited in common by both genera viral integration were shown to terminate in palindromic of PDVs may be similar manifestations resulting from sequences of 4-base pair length, “GATC”. These palin- common viral life cycles, or similarly evolved viral rec- dromic LTR termini sequences were then suggested to ognition mechanisms. be sites for recombination of the segment W to form We have described a putative in vitro integration nested segments R and M. The direct repeats present on mechanism for one segment of the GiPDV DNA gen- proviral termini of bracoviruses identified in CiV and ome. It is necessary to examine the other GiPDV gen- CcPDV have been much shorter, from 14 to 24 base ome segments for which we have evidence of mainte- pairs (Savary et al., 1997; Wyder et al., 2002), as we nance in transformed insect cells to further develop a would similarly expect for GiPDV. We may predict the model for in vitro integration and propagation of PDVs. sequences “CATG” and its palindrome “GTAC” may In addition, we would like to examine GiPDV in vitro represent short repeated sequences from the site of integration in numerous other infected insect cell lines GiPDV segment integration in the G. indiensis parasit- derived from a variety of insect hosts. Stable integration oid. The excision of the segment F from the parasitoid of exogenous GiPDV sequences into the insect cell genome at these sites, followed by circularization to nuclear genome in vitro is highly significant for its form the circular segment F, could result in their joining potential to be used for continuous transformation of to form the “CATGGTAC” seen in the circular segment insect cells. For the future, it will be essential to investi- F. Thus the “CATGGTAC” palindromic repeat present gate whether or not PDV sequences may be similarly in viral sequences directly at the site of in vitro inte- integrated in insect cells in vivo. gration would represent a recognition site for viral integration by virtue of structure and sequence. In vitro lin- earization and integration of segment F would directly mirror the integrated proviral state of the integration in Acknowledgements the parasitoid in vivo. However, we must examine GiPDV segment F integrated in the parasitoid host as The authors gratefully acknowledge USDA NRI grant provirus to test this prediction. 99-35302-8522 for supporting this study, Curtis Huber From an evolutionary standpoint, a virus such as PDV and Frank Washington III for excellent technical assist- with an integrated provirus-excision-circularization life- ance, and Philip B. Taylor (USDA, Beneficial Insects style might be expected to also integrate segments in Research Laboratory, Newark, DE) for rearing the G. host cells and that these could be maintained in similar indiensis parasitoid wasps. D.E. Gundersen-Rindal, D.E. Lynn / Journal of Insect Physiology 49 (2003) 453–462 461

References and adult stages of the female. Journal of General Virology 77, 2873–2879. Abe, H., Ohbayashi, F., Sugasaki, T., Kanehara, M., Terada, T., Shim- Gundersen-Rindal, D., Dougherty, E.M., 2000. Evidence for inte- ada, T., Kawai, S., Mita, K., Kanamori, Y., Yamamoto, M.T., gration of Glyptapanteles indiensis polydnavirus DNA into the Oshiki, T., 2001. Two novel Pao-like retrotransposons (Kamikaze chromosome of Lymantria dispar in vitro. Virus Research 66, and Yamato) from the silkworm species Bombyx mori and B. man- 27–37. darina: common structural features of Pao-like elements. Molecular Gundersen-Rindal, D., Lynn, D.E., Dougherty, E.M., 1999. Transform- Genetics and Genomics 265, 375–385. ation of lepidopteran and coleopteran insect cell lines by Glypta- Albrecht, U., Wyler, T., Pfister-Wilhelm, R., Gruber, A., Stettler, P., panteles indiensis polydnavirus DNA. In Vitro Cellular and Devel- Heiniger, P., Kurt, E., Schumperli, D., Lanzrein, B., 1994. Polydna- opmental Biology 35, 111–114. virus of the parasitic wasp Chelonus inanitus (Braconidae): charac- Hutchison, C.A., Hardies, S.C., Loeb, D.D., Shehee, W.R., Edgell, terization, genome organization and time point of replication. Jour- M.H., 1989. LINES and related retroposons: Long interspersed nal of General Virology 75, 3353–3363. repeated sequences in the eukaryotic genome. In: Berg, D.E., Altschul, S.F., Gish, W., Miller, W., Myers, E.W., Lipman, D.J., 1990. Howe, M.M. (Eds.), Mobile DNA. American Society for Micro- Basic local alignment search tool. Journal of Molecular Biology biology, Washington, DC, pp. 594–617. 215, 403–410. Jakubczak, J.L., Xiong, Y., Eickbush, T.H., 1990. Type I (R1) and Altschul, S.F., Madden, T.L., Scha¨ffer, A.A., Zhang, J., Zhang, Z., type II (R2) ribosomal DNA insertions of Drosophila melanogaster Miller, W., Lipman, D.J., 1997. Gapped BLAST and PSI-BLAST: are retrotransposable elements closely related to those of Bombyx a new generation of protein database search programs. Nucleic mori. Journal of Molecular Biology 212, 37–52. Acids Research 25, 3389–3402. Johnson, M.S., McClure, M.A., Feng, D.F., Gray, J., Doolittle, R.F., Asgari, S., Schmidt, O., Theopold, U., 1997. A polydnavirus-encoded 1986. Computer analysis of retroviral pol genes: assignment of protein of an endoparasitoid wasp is an immune suppresser. Journal enzymatic functions to specific sequences and homologies with of General Virology 78, 3061–3070. nonviral enzymes. Proceedings of the National Academy of Ausubel, F.M., Brent, R., Kingston, R.E., Moore, D.D., Seidman, J.G., Sciences 83, 7648–7652. Smith, J.H., Struhl, K., 1994. Current Protocols in Molecular Kim, M.-K., Sisson, G., Stoltz, D.B., 1996. Ichnovirus infection of an Biology, vol. 1, chapter 2. John Wiley and Sons, New York. established gypsy moth cell line. Journal of General Virology 77, Beckage, N.E., Tan, F.F., Schleifer, K.W., Lane, R.D., Cherubin, L., 2321–2328. 1994. Characterization and biological effects of Cotesia congregata Lawrence, P.O., Lanzrein, B., 1993. Hormonal interactions between polydnavirus on host larvae of the tobacco hornworm, Manduca insect endoparasites and their host insects. In: Beckage, N.E., sexta. Archives of Insect Biochemistry and Physiology 26, 165– Thompson, S.N., Federici, B.A. (Eds.), Parasites and Pathogens of 195. Insects, Vol. 1. Academic Press, San Diego, CA, pp. 59–86. Blissard, G.W., Vinson, S.B., Summers, M.D., 1986. Identification, Lavine, M.D., Beckage, N.E., 1995. Polydnaviruses: potent mediators mapping, and in vitro translation of Campoletis sonorensis virus of host insect immune dysfunction. Parasitology Today 11, 368– mRNAs from parasitized Heliothis virescens larvae. Journal of 378. Virology 57, 318–327. McKelvey, T.A., Lynn, D.E., Gundersen-Rindal, D., Guzo, D., Stoltz, Chen, Y.P., Gundersen-Rindal, D.E., 2003. Morphological and gen- D.A., Guthrie, K.P., Taylor, P.B., Dougherty, E.M., 1996. Trans- omic characterization of the polydnavirus associated with the para- formation of gypsy moth (Lymantria dispar) cell lines by infection sitoid wasp Glyptapanteles indiensis (Hymenoptera: Braconidae). with Glyptapanteles indiensis polydnavirus. Biochemical and Journal of General Virology 84, in press. Biophysical Research Communications 225, 764–770. Cui, L., Webb, B.A., 1997. Homologous sequences in the Campoletis Savary, S., Beckage, N., Tan, F., Periquet, G., Drezen, J.-M., 1997. sonorensis polydnavirus genome are implicated in replication and Excision of the polydnavirus chromosomal integrated EP1 nesting of the W segment family. Journal of Virology 71, 8504– sequence of the parasitoid wasp Cotesia congregata (Braconidae, 8513. Microgastinae) at potential recombinase binding sites. Journal of Eickbush, T.H., 1992. Transposing without ends: the non-LTR retro- General Virology 78, 3125–3134. transposable elements. The New Biologist 4, 430–440. Southern, E.M., 1975. Gel electrophoresis of restriction fragments. Fleming, J.G.W., 1991. The integration of polydnavirus genomes in Journal of Molecular Biology 98, 503–517. parasitoid genomes: implications for biocontrol and genetic analy- Stoltz, D.B., 1990. Evidence for chromosomal transmission of polyd- ses of parasitoid wasps. Biological Control 1, 127–135. navirus DNA. Journal of General Virology 71 (1051-105), 6. Fleming, J.G.W., Summers, M.D., 1986. Campoletis sonorensis endo- Stoltz, D.B., 1993. The polydnavirus life cycle. In: Beckage, N.E., parasitic wasps contain forms of C. sonorensis virus DNA sugges- Thompson, S.N., Federici, B.A. (Eds.), Parasites and Pathogens of tive of integrated and extrachromosomal polydnavirus DNAs. Jour- Insects, Vol. 1. Academic Press, San Diego, CA, pp. 167–187. nal of Virology 57, 552–562. Stoltz, D., Whitfield, J.B., 1992. Viruses and virus-like entities in the Fleming, J.G.W., Summers, M.D., 1990. The integration of the genome parasitic hymenoptera. Journal of Hymenoptera Research 1, 125– of a segmented DNA virus in the host insect’s genome. In: Hage- 139. dorn, H.H., Hildebrand, J.G., Kidwell, M.G., Law, J.H. (Eds.), Stoltz, D.B., Guzo, D., Cook, D., 1986. Studies on polydnavirus trans- Molecular Insect Science. Plenum Press, New York, pp. 99–105. mission. Virology 155 (120-13), 1. Fleming, J.G.W., Summers, M.D., 1991. Polydnavirus DNA is inte- Stoltz, D.B., Beckage, N.E., Blissard, G.W., Fleming, J.G.W., Krell, grated in the DNA of its parasitoid wasp host. Proceedings of the P.J., Theilmann, D.A., Summers, M.D., Webb, B.A., 1995. Polyd- National Academy of Sciences 88, 9770–9774. naviridae. In: Murphy, F.A., Fauquet, C.M., Bishop, D.H.L., Gabriel, A., Boeke, J.D., 1993. Retrotransposon reverse transcription. Ghabrial, S.A., Jarvis, A.W., Martelli, G.P., Mayo, M.A., Sum- In: Skalka, A.M., Goff, S.P. (Eds.), Reverse Transcriptase. Cold mers, M.D. (Eds.), Virus Taxonomy. Sixth report of the inter- Spring Harbor Laboratory Press, Cold Spring Harbor, pp. 275–328. national committee on taxonomy of viruses. Springer Verlag, Grandgenett, D.P., 1986. Nuclease mechanism of the avian retrovirus Vienna, pp. 143–147. pp32 endonuclease. Journal of Virology 58, 970–974. Strand, M.R., McKenzie, D.I., Grassl, V., Dover, B.A., Aiken, J.M., Gruber, A., Stettler, P., Heiniger, P., Schumperli, D., Lanzrein, B., 1992. Persistence and expression of Micropletis demolitor polydna- 1996. Polydnavirus DNA of the braconid wasp Chelonus inanitus virus in Pseudoplusia includens. Journal of General Virology 73, is integrated in the wasp’s genome and excised only in later pupal 1627–1635. 462 D.E. Gundersen-Rindal, D.E. Lynn / Journal of Insect Physiology 49 (2003) 453–462

Strand, M.R., Pech, L.L., 1995. Immunological basis for compatibility Volkoff, A.-N., Rocher, J., Cerutti, P., Ohresser, M.C.P., d’Aubenton- in parasitoid-host relationships. Annual Reviews of Entomology Carafa, Y., Devauchelle, G., Duonor-Cerutti, M., 2001. Persistent 40, 31–56. expression of a newly characterized Hyposoter didymator polydna- Summers, M.D., Dibb-Hajj, S., 1995. Polydnavirus-facilitated endo- virus gene in long-term infected lepidopteran cell lines. Journal of parasite protection against host immune defenses. Proceedings of General Virology 82, 963–969. the National Academy of Sciences 92, 29–36. Webb, B.A., 1998. Polydnavirus biology, genome structure, and evol- Takahashi, H., Okazaki, S., Fujiwara, H., 1997. A new family of site- ution. In: Miller, L.K., Ball, L.A. (Eds.), The Insect Viruses. Ple- specific retrotransposons, SART1, is inserted into telomeric repeats num Publishing Corporation, New York, pp. 105–139. of the silkworm, Bombyx mori. Nucleic Acids Research 25, Whitfield, J.B., 1997. Molecular and morphological data suggest a sin- 1578–1584. gle origin of the polydnaviruses among braconid wasps. Naturwis- Theilmann, D.A., Summers, M.D., 1986. Molecular analysis of Cam- senschaften 84, 502–507. poletis sonorensis virus DNA in the lepidopteran host Heliothis Whitfield, J.B., 2002. Estimating the age of the polydnavirus braconid virescens. Journal General Virology 67, 1961–1969. wasp symbiosis. Proceedings of the National Academy of Sciences Theilmann, D.A., Summers, M.D., 1987. Physical analysis of the Cam- 99, 7508–7513. poletis sonorensis virus multipartite genome and identification of Wyder, S., Tschannen, Hochuli, A., Saladin, V., Zumbach, S., Lanzr- a family of tandemly repeated elements. Journal of Virology 61, ein, B., 2002. Characterization of Chelonus inanitus polydnavirus 2589–2598. segment sequences and analysis, excision site, and demonstration Volkoff, A.-N., Cerutti, P., Rocher, J., Ohresser, M.C.P., Devauchelle, of clustering. Journal of General Virology 83, 247–256. G., Duonor-Cerutti, M., 1999. Related RNAs in Lepidopteran cells Xu, D., Stoltz, D., 1991. Evidence for a chromosomal location of after in vitro infection with Hyposoter didymator virus define a polydnavirus DNA in the ichneumonid parasitoid Hyposoter fugi- new polydnavirus gene family. Virology 263, 349–363. tivus. Journal of Virology 65, 6693–6704.